基于PCA-LVQ的专业可持续发展综合分类研究

Study on Comprehensive Classification of the Sustainable Development ofProfession Based on PCA-LVQ

提出将PCA及LVQ网络用于专业可持续发展分类应用研究,通过多种调研形式获取专业发展的21项指标,首先对比使用标准正交归一化方法和min-max归一化方法对原始的样本进行归一化,再利用PCA算法解除样本特征的相关性,实现数据的降维。最后利用LVQ神经网络对已经降维的数据进行学习,并得到学习结果,最后利用学习得到的权值矩阵对新的样本进行识别,判断专业发展水平。

In this paper, PCA-LVQ networks are applied to the classification and application ofthe sustainable development of profession and 21 indexes of professional development are obtained through various investiga- tion forms. First, the standard orthogonal normalization method and the min-max normalization method are compared to normalize the original samples. Then the PCA algorithm is used to remove the correlation of the sample features and to reduce the dimension of thedata. Finally, LVQ neural network is used to study the dimension reduced data and the learning result is obtained. The weight matrix is used to identify the new samples and judge the level of professional development.